alphabet’s google is in talks with chip design firm marvell technology to jointly develop two new chips specifically engineered to efficiently run artificial intelligence models. this move is seen as a significant step by google to further challenge nvidia’s dominant position in the ai chip market.
according to reports, the two chips will each serve a specific purpose. one is a memory processing unit designed to work in tandem with google’s self-developed tensor processing unit (tpu), while the other is a brand-new tpu built from the ground up for running ai models. for years, google has been striving to establish its tpu as a viable alternative to nvidia’s mainstream graphics processing units (gpus). as google seeks to demonstrate to investors that its ai investments are paying off, tpu sales have become a key driver of revenue growth for google cloud.
sources indicate that the two companies plan to complete the design of the memory processing unit as early as next year, followed by pilot production. if the collaboration proceeds smoothly, google could further reduce its reliance on nvidia in the ai hardware space while strengthening the competitiveness of its cloud computing services. at present, neither party has issued an official statement on the matter; however, industry observers generally believe that this development signals intensifying competition in the ai chip market.
